PYL9 is involved in the regulation of ABA signaling during tomato fruit ripening

Abstract

Abscisic acid (ABA) regulates fruit ripening, yet little is known about the exact roles of ABA receptors in fruit. In this study, we reveal the role of SlPYL9, a tomato pyrabactin resistance (PYR)/pyrobactin resistance-like (PYL)/regulatory component of ABA receptors (RCAR) protein, as a positive regulator of ABA signaling and fruit ripening. SlPYL9 inhibits protein phosphatase-type 2C (PP2C2/6) in an ABA dose-dependent way, and it interacts physically with SlPP2C2/3/4/5 in an ABA-dependent manner. Expression of SlPYL9 was observed in the seeds, flowers, and fruits. Overexpression and suppression of SlPYL9 induced a variety of phenotypes via altered expression of ABA signaling genes (SlPP2C1/2/9, SlSnRK2.8, SlABF2), thereby affecting expression of ripening-related genes involved in ethylene release and cell wall modification. SlPYL9-OE/RNAi plants showed a typical ABA hyper-/hypo-sensitive phenotype in terms of seed germination, primary root growth, and response to drought. Fruit ripening was significantly accelerated in SlPYL9-OE by 5–7 d as a result of increased endogenous ABA accumulation and advanced release of ethylene compared with the wild-type. In the SlPYL9-RNAi lines, fruit ripening was delayed, mesocarp thickness was enhanced, and petal abscission was delayed compared with the wild-type, resulting in conical/oblong and gourd-shaped fruits. These results suggest that SlPYL9 is involved in ABA signaling, thereby playing a role in the regulation of flower abscission and fruit ripening in tomato.

ABA receptor PYL9 plays a role in the regulation of flower abscission and fruit ripening in tomato by modulating ABA signaling components, which regulates expression of related genes.

Introduction

Abscisic acid (ABA) is an essential endogenous messenger during plant development and stress responses (Wasilewska et al., 2008; Peleg and Blumwald, 2011). The ABA-mediated signaling cascade is initiated by the perception of ABA by ABA receptors, a large family of soluble pyrabactin resistance (PYR)/pyrobactin resistance-like (PYL)/regulatory component of ABA receptors (RCAR) proteins (Ma et al., 2009; Park et al., 2009). PYL-type 2C phosphatase (PP2C) interaction releases SNF1-related kinase 2 (SnRK2) downstream (Fujii et al., 2009; Umezawa et al., 2009), causing phosphorylation of downstream proteins and activation of ion channels (Hubbard et al., 2010; Weiner et al., 2010). ABA receptors can be grouped into three subfamilies based on sequence homology, the oligomeric state in solution, ABA sensitivity, and basal activation levels (Ma et al., 2009; Yin et al., 2009). Subfamilies I and II are monomeric, with high basal activity, and require only a low level of ABA to induce PP2C inhibition, while subfamily III receptors are dimeric in solution and have low basal activity (Santiago et al., 2009).

Increasing evidence from gene expression, protein activity, and β-glucuronidase (GUS) promoter fusion analyses suggests that ABA receptors differ spatiotemporally in mRNA and protein accumulation (González-Guzmán et al., 2012, 2014). During stress and according to developmental requirements, ABA levels increase several-fold (Zeevaart and Creelman, 1988), leading to a broad array of tissue- and context-specific physiological responses. It is therefore possible that different physiological responses are induced by different ABA receptors. Moreover, it has also been suggested that the intensity of activation during these physiological responses is dependent on receptor sensitivity rather than on the quantity of ABA (Santiago et al., 2009; Okamoto et al., 2013).

Recently, the role of ABA in regulating both climacteric and non-climacteric fruit ripening was revealed (Galpaz et al., 2008; Romero et al., 2012). In tomato fruit, exogenous ABA induces ethylene biosynthesis via regulation of ACS (encoding 1-aminocyclopropane-1-carboxylic acid (ACC) synthase) and ACO (encoding ACC oxidase) gene expression (Zhang et al., 2009). In non-climacteric fruit such as strawberry (Jia et al., 2011; Li et al., 2013) and grape (Sun et al, 2010), exogenous ABA improves color. Moreover, in ABA-deficient tomato mutants, the fruit shows an abnormal growth pattern compared with the wild-type (WT) (Galpaz et al., 2008). This is accompanied by a decrease in expression of several genes encoding ripening-associated cell wall enzymes, resulting in altered fruit texture and coloration (Sun et al., 2012a,b; Ji et al., 2014). A role of ABA during fruit set in tomato has also been implicated (McAtee et al., 2013), while ABA glucose conjugation mediated by uridine diphosphate glucosyltransferases was found to be an important pathway in regulation of ABA homeostasis in tomato fruit (Sun et al., 2017). In addition, transcriptional factors have also been shown to play a role in ABA-mediated fruit ripening, e.g. VvABF2 in grape (Nicolas et al., 2014), PacMYBA in sweet cherry (Shen et al., 2014), and the zinc finger transcription factor SlZFP2 in tomato (Weng et al., 2015). Furthermore, chemical screening of ABA agonists using Arabidopsis thaliana receptors revealed small molecules that activate ABA signaling pathways across different plant species (Okamoto et al., 2013).

However, despite emerging research (Pilati et al., 2017; Sun et al., 2017; Zhang et al., 2018), little is known about the role of core ABA signaling components in fruit. Understanding the role of each PYL ABA receptor in plant development is very important in both basic science research and applied technology. In this study, we therefore evaluated the physiological roles of the ABA receptor SlPYL9 during fruit ripening in tomato using overexpression (OE) and RNAi-mediated transgenics. Accordingly, SlPYL9 was found to play a role in ABA-mediated fruit ripening in tomato.

Materials and methods

Generation of SlPYL9 transgenic tomato lines

Using the PCR primers shown in Supplementary Table S1, available at JXB online, the 35S::antisense SlPYL9::GUS::sense SlPYL9::Nos-terminator fusion gene or 35S::SlPYL9::Nos-terminator fusion gene was cloned into pCAMBIA1305.1 (Invitrogen, Carlsbad, CA, USA), and the resulting RNAi/OE constructs introduced into Micro-Tom tomato via Agrobacterium tumefaciens LBA4404-mediated transformation. Positive transgenic plants were identified through GUS staining. The T₁ generation, showing a 3:1 segregation ratio of transgenic and non-transgenic plants, was considered representative of a single copy insertion line. T₂ seeds were collected from each T₁ plant, with T₂ plants not showing segregation and thought to represent homozygosis lines. All analyses were conducted using T₂ homozygous plants of OE/RNAi lines with the non-transgenic WT as a control.

Plant materials and gene cloning

Tomato plants (Solanum lycopersicum L. cv. Micro-Tom) of WT, SlPYL9-OE, and RNAi lines were grown under standard greenhouse conditions (25±5 °C, 70% humidity, 14/10 h light/dark regime, 30–50 plants per line). Fruit ripening stages were determined according to the number of days after full bloom (DAFB) and fruit color as follows (for WT fruit): immature green (IM), 20, 24, and 26 DAFB; mature green (MG), 28 DAFB; breaker (B), 32 DAFB; turning (T), 34 DAFB; red (R), 40 DAFB; and over-ripe (OR), 42 DAFB. Fifteen fruits were harvested randomly at each stage from each transgenic line and WT plants then divided into three groups as three replicates. After harvest, each group of fruits was weighed and ethylene production investigated. All fruits were then dissected into three parts (peel, pulp, and seed), frozen in liquid nitrogen, and stored at –80 °C until further use.

ABA and nordihydroguaiaretic acid treatment

To determine the effect of ABA and nordihydroguaiaretic acid (NDGA) treatment on ripening onset of fruits still attached to the plant, 150 fruits from WT, two SlPYL9-OE, and two SlPYL9-RNAi lines were selected at the MG stage, and divided into three groups; 0.5 ml of ABA (100 μM) was injected into group I fruit, 0.5 ml of NDGA (200 μM) into group II fruit, and distilled water into group III fruit as a control. After 0, 2, 4, 6, and 8 d, the fruits were sampled for further analysis.

qRT-PCR analysis

RNA extraction was performed using the SV Total RNA Isolation System (Promega, Madison, WI, USA), and the RNA was subsequently digested with DNase I (Takara). The extracted RNA was reverse transcribed to cDNA using the Takara RNA PCR Kit (Takara, Kyoto, Japan). Quantitative real-time PCR (qRT-PCR) detection was performed using SYBR Premix ExTaq (Perfect Real Time; Takara Bio) on a Rotor-Gene 3000 system (Corbett Research, Sydney, NSW, Australia) to quantify gene expression levels. SAND (SGN-U316474) was employed as an internal control (Expósito-Rodríguez et al., 2008). The expression level of each gene was calculated using Rotor-Gene 6.1.81 software with two standard curves. The primer sequences are shown in Supplementary Tables S2, S3.

ABA receptor activity assays in vitro

PP2C and PYL proteins were expressed and purified as described previously (Okamoto et al., 2013), with minor modifications. The ΔN-PP2C2 (aa 96–409) and ΔN-PP2C6 (aa 176–508) cDNAs were ligated into the pET-28a vector, which was transformed into Escherichia coli strain BL21 (DE3). Escherichia coli DE3 cells were grown to an OD₆₀₀ of 0.8–0.9, and then 1 mM isopropyl β-D-1-thiogalactopyranoside was added, and the cells were grown at 23 °C for 16 h with shaking to induce target protein expression. The proteins were then purified using Ni-NTA His·Bind Resin (Novagen) as described previously (Yamaguchi-Shinozaki and Shinozaki, 2006). Similarly, PYL9 protein was obtained as described previously (Nambara et al., 2005). To analyse the inhibition of PYL9 on PP2C2/6, the purified proteins (final concentration: 200 nM of PP2C2/6, 600 nM of PYL9) were added into 80 μl of assay buffer that contained 10 mM MnCl₂ and 0.1% β-mercaptoethanol with or without ABA. The reactions took place after 20 μl of reaction solution was added, which contained 250 mM p-nitrophenyl phosphate (pNPP) substrate, 165 mM Tris-acetate (pH 7.9), 330 mM potassium acetate, and 0.5% BSA, and the system was incubated at 22 °C for 30 min. The end products from the reaction were detected through a Multiskan Mk3 microplate reader (Thermo Fisher Scientific) with a detection wavelength of 405 nm.

To analyse SlPP2C2/6 enzyme activity, 1 μg of ΔN-PP2C2/6 protein was added to 80 μl reaction mixtures with 10 mM MnCl₂ and 0.1% β-mercaptoethanol. The reaction took place after addition of 20 μl of reaction solution that contained 250 mM pNPP, 165 mM Tris-acetate (pH 7.9), 330 mM potassium acetate and 0.5% BSA. During the 30 min incubation at 22 °C, the products were detected at each 5 min interval. The product amount was determined through a standard curve of 4-nitrophenol made in the same buffer system. The reaction progressions were plotted and the ΔN-PP2C2/6 enzyme activity was calculated (Park et al., 2009).

Yeast two-hybrid assay

To find the interactions among SlPYL9–SlPP2Cs, an SlPYL9 gene fragment with restriction sites was cloned into the pGADT7 vector while SlPP2Cs was cloned into the pGBKT7 vector, and they were then co-transformed into yeast. The interactions were determined by a growth assay on medium lacking Leu and Trp (as control), and on media lacking Leu, Trp, His, and Ade with ABA in different concentrations.

Subcellular localization and bimolecular fluorescence complementation assay

To investigate subcellular localization, the coding open reading frame (ORF) sequences of SlPYL9 tagged with green fluorescent protein (GFP) at the C-terminus were cloned into the pCambia1300 vector. The specific primers are listed in Supplementary Table S4. Agrobacterium tumefaciens GV3101 carrying pCambia1300-SlPYL9-GFP was then injected into 5- to 6-week-old Nicotiana benthamiana leaves. After 2 d culture in the dark, the infiltrated leaves were examined, and GFP signals monitored using a fluorescence microscope (Olympus BX51, Japan) with an optical wavelength of 470 nm at ×200 magnification.

For bimolecular fluorescence complementation (BiFC), SlPYL9 was cloned into pCambia1300-35S-YFPN and SlPP2Cs into the pCambia1300-35S-YFPC (Waadt et al., 2008) using the primers listed in Supplementary Table S5. Agrobacterium tumefaciens GV3101 carrying the fusion constructs was then co-infiltrated into 5- to 6-week-old N. benthamiana leaves. After 24 h, leaves were sprayed with ABA. Fluorescence signals were visualized 4 d after infiltration under a fluorescence microscope (Olympus BX51, Japan).

Measurement of ABA

Three grams of pulp tissues was extracted with 40 ml of 80% (v/v) methanol at –20 °C for 18 h. The methanol extracts were then centrifuged at 10 000 g for 20 min, and the pellet was extracted twice with 20 ml of 80% methanol at –20 °C for 2 h. The supernatants were subsequently combined and dried under vacuum, and the residue was dissolved in 10 ml of 0.02 M phosphate assay buffer (pH 8.0) and extracted three times with 10 ml petroleum ether. The organic phase was removed, and the pH of the aqueous phase was adjusted to 8.0, followed by the addition of 0.2 g of insoluble polyvinylpolypyrrolidone. After stirring for 15 min at 0 °C, polyvinylpolypyrrolidone was removed through vacuum filtration. The mixed liquid was adjusted to pH 3.0 and subsequently extracted three times with ethyl acetate. The ethyl acetate phase was dried under vacuum and dissolved in 1 ml 50% methanol (v/v). The ABA content was determined via HPLC (Agilent Technologies 1200) using a 4.8×150 mm C18 column (Agilent Technologies), at a flow rate of 0.8 ml min⁻¹. Elution was performed using both solvent A (0.8% (v/v) glacial acetic acid) and solvent B (100% methanol). We employed (±)-abscisic acid (Sigma-Aldrich, St Louis, MO, USA) as the standards for determination at 260 nm. Three replicates were conducted for each sample.

Determination of ethylene production

Ethylene production was measured by enclosing three fruits in 50 ml airtight containers for 2 h at 20 °C, then withdrawing 1 ml of the headspace gas and injecting it into a gas chromatograph (Agilent model 6890N) fitted with a flame ionization detector and an activated alumina column. Fresh tissues from each fruit were frozen in liquid nitrogen and stored at −80 °C until further use.

In situ hybridization

Tissue fixation and in situ hybridizations were performed as described previously (Hord et al., 2006) with minor modifications. Digoxigenin-labelled in situ probes were synthesized by PCR amplification of cDNA using gene-specific primers containing T7 and SP6 RNA polymerase binding sites. Antisense probes were generated using T7 RNA polymerase, and sense probes constructed using SP6 RNA polymerase. The gene primer pairs are listed in Supplementary Table S6.

RNA-Seq

Total RNA was extracted from breaker fruits, then a total of 3 μg RNA per sample was used for mRNA purification and library construction using the Truseq™ RNA Sample Prep Kit (Illumina, CA, USA). The samples were sequenced on an Illumina HiSeq™ 2000. Each sample yielded more than 6 G of data. Clean data were obtained by removing reads from raw data. The clean reads were aligned to the reference tomato genome release SL2.50 using TopHat v2.0.13 (Kim et al., 2013). HTSeq v0.5.3 was used to count the number of reads mapped to each gene, then the RPKM of each gene was calculated based on the length of the gene and reads count mapped to this gene. Differential expression analysis was performed using the DEGSeq R package (1.12.0) and Cufflinks (Trapnell et al., 2012). A corrected P-value of 0.005 and log2 (fold change) of 1 were set as the threshold for significantly differential expression.

Statistical treatment of the data

Data were statistically analysed by SPSS software using one-way analysis of variance, and Duncan’s test of significance. Levels of significance *P<0.05 and **P<0.01 were determined by t-test.

Results

Expression of SlPYL9 in seeds, fruits, and flowers during their development

In silico analysis of the tomato genome revealed that there are 14 SlPYL putative ABA receptors, all of which shared highly conserved elements with Arabidopsis (see Supplementary Fig. S1). They were subsequently clustered into three subfamilies based on phylogenetic analysis (González-Guzmán et al., 2014). qRT-PCR results showed that among the 14 SlPYLs, the expression level of SlPYL1 was the highest, and the expression patterns of SlPYL1, SlPYL4, and SlPYL9 agreed with the ABA accumulation in fruits during development (Supplementary Fig. S2). The SlPYL9 protein was closest to monomeric AtPYL4 (Supplementary Fig. S1), which belongs to subgroup II in the Arabidopsis genome (Dupeux et al., 2011). qRT-PCR results showed that the expression of SlPYL9 was the highest in the seeds, and it was also expressed in flowers, leaves, roots, and stems (Fig. 1A). In mesocarp, the expression of SlPYL9 rapidly increased and peaked at the breaker stage, with the concomitant onset of coloration, after which it declined before increasing once again at the B+10 stage (Fig. 1A). In situ hybridization of flower buds and young fruits was subsequently performed to further examine SlPYL9 expression. SlPYL9 was highly expressed in the stigma, ovaries, and anthers 3 d before flowering (Fig. 1B), after which it was highly expressed in the seeds, and less so in the fruit peel/pulp, vascular bundle, and connective placental regions, at 6 d after full bloom (Fig.1C). Our result was consistent with a previous report (González-Guzmán et al., 2014). These findings suggest that SlPYL9 plays a role in the development of seed, fruit, flower, and vegetative organs such as roots.

Fig. 1.

SlPYL9 expression patterns in Micro-Tom tomato. (A) Expression levels of SlPYL9 in seeds, fruits, flowers (Fl), leaves (Le), roots (Ro), and stems (St). (B, C) Spatio-temporal expression of SlPYL9 in floral buds (B) and young fruits (C) determined through in situ hybridization. Data are based on three independent replicates. Error bars indicate the SD (n=3). (This figure is available in color at JXB online.)

SlPYL9-dependent SlPP2C2/6 inhibition and interaction between SlPYL9 and SlPP2Cs

First, we verified whether SlPYL9 is an ABA receptor. To do so, SlPYL9-mediated SlPP2C2/6 inhibition was examined in vitro. The concentration of SlPP2C2/6 was 200 nM and the concentration of SlPYL9 was 3-fold higher. In this case, SlPP2C2 and SlPP2C6 were inhibited by 54% and 61%, respectively, by SlPYL9 protein in 10 μM ABA (Fig. 2A–C). The enzyme activities of SlPP2C2 and SlPP2C6 were 0.5433 and 1.2189 µmol min⁻¹ μg⁻¹, respectively (Fig. 2D, E). Subsequently, analysis of interactions of SlPYL9 with SlPP2Cs was carried out, using both the yeast two-hybrid (Y2H) method and a BiFC assay. The Y2H results showed that SlPYL9 is able to respectively interact with SlPP2C2/3/4/5, and furthermore, these interactions were enhanced with increasing concentrations of ABA (Fig. 3A). In contrast, no interaction with SlPP2C1 was observed except at very high concentrations of ABA. Next, tagged SlPYL9 and SlPP2Cs were transiently expressed in N. benthamiana. As shown in Fig. 3B, stronger fluorescence was observed in the cytoplasm with SlPYL9–YFPN and SlPP2C4/5–YFPC combinations, and in the cell nucleus and cytoplasm with SlPYL9–SlPP2C3 in an ABA-dependent manner. In the above Y2H, BiFC, and SlPYL9-mediated SlPP2Cs inhibition tests, the results were not entirely consistent in the different experimental systems, and the reasons need to be further studied. The results of subcellular localization are shown in Fig. 3C. The results of GFP-tagged fluorescence revealed that SlPYL9 is localized in the nucleus and cytoplasm, suggesting specific involvement in ABA signaling.

Fig. 2.

SlPYL9 inhibits SlPP2C2/6 activity in the presence of ABA. (A) Prokaryotic expression and purification of recombinant SlPYL9 and SlPP2C2/6. M, protein molecular marker; 1, no induction of SlPYL9; 4 and 7, no-induction of SlPP2C 2/6; 2, 5, and 8, induction of SlPYL9 and SlPP2C2/6; 3, 6, and 9, purification of SlPYL9 and SlPP2C2/6 proteins; arrows indicates target proteins. (B, C) Inhibition of SlPP2C2/6 enzyme activity by SlPYL9 in response to ABA. SlPP2C2/6 activity in absence of ABA and presence of receptor is shown as 100% activity. (D, E) Determination of the SlPP2C2/6 enzyme activity as the amount of 4-nitrophenol produced in µmol per min of 1 µg PP2C. (This figure is available in color at JXB online.)

Fig. 3.

Interaction between SlPYL9 and SlPP2Cs. (A) Yeast two-hybrid (Y2H) assay between SlPYL9 and SlPP2Cs. (B) A bimolecular fluorescence complementation (BiFC) assay of SlPYL9 and SlPP2Cs. (C) Subcellular localization of SlPYL9. (This figure is available in color at JXB online.)

Overexpression and suppression of SlPYL9 affects expression of ABA-responsive genes

To further understand the exact function of SlPYL9, we generated overexpression and RNAi lines to respectively promote and reduce levels of SlPYL9 (see Supplementary Fig. S3). A total of five independent SlPYL9-OE and four SlPYL9-RNAi transgenic lines were created, and of these, two OE and three RNAi independent T₂ homozygous single insertion lines were selected. RNA-Seq analyses of mesocarp (breaker stage) showed that expression of SlPYL9 was significantly up-regulated in OE fruit and reduced in RNAi lines (Fig. 4A). Altered expression of SlPYL9 had little effect on the expression of other PYL family members (Fig. 4A). Among the SlPP2C genes, most were up-regulated in both OE and RNAi fruits (Fig. 4B). For the SnRK2s, only SlSnRK2.8 was significantly down-regulated in OE fruits and up-regulated in RNAi fruits, while the expression of other SnRK2s was up-regulated in RNAi fruits and showed no obvious changes in OE fruits (Fig. 4C). Up-regulation of only one member (SlAREB1/ABF2) of the ABA responsive element binding factor subgroup, AREB/ABF, which mediates ABA-responsive gene transcription, was observed at the breaker stage (Fig. 4D). SlAREB1/ABF2 activated down-stream genes of the ABA signaling cascade, down/up-regulating the expression of ripening-related transcription factor CNR in the SlPYL9-OE/RNAi fruit at the breaker stage (Fig. 4D). As shown Fig. 4D, among fruit ripening regulators, the expression of CNR was significantly down-regulated in OE fruits but was up-regulated in RNAi fruits. The expression of RIPENING-INHIBITOR (RIN) and TOMATO AGAMOUS LIKE 1 (TAGL1) was up-regulated in RNAi fruits but there was no significant difference between OE and WT fruits. The expression of MADS-BOX TRANSCRIPTION FACTOR 1 (MADS1) showed no obvious change compared with WT fruits. Collectively, these results suggest that SlPYL9 is involved in the regulation of ABA signaling in Micro-Tom tomato during fruit ripening.

Fig. 4.

Relative expression levels of genes related to ABA signaling and transcriptional factors in WT and SlPYL9-OE/RNAi transgenic fruit. Gene expression was examined at the breaker stage using RNA-Seq analysis.

Sensitivity of SlPYL9-OE/RNAi transgenic tomato to ABA

Sensitivity of the transgenic lines to ABA was evaluated by measuring seed germination rates, primary root growth, and drought resistance. Null segregating transgenic plants were used as a reference for the basal ABA response. As a result, significantly enhanced/reduced ABA sensitivity was observed in seeds of OE/RNAi plants. This resulted in delayed germination and primary root growth in the OE lines, and opposite effects in the RNAi lines (see Supplementary Fig. S4). In addition, increased drought resistance was observed in the OE lines, while resistance was weak in the RNAi lines compared with the WT (Supplementary Fig. S5). We also assayed the effects of exogenous ABA and nordihydroguaiaretic acid (NDGA), an inhibitor of ABA synthesis, on fruit ripening onset. MG fruits attached to the plant were treated with ABA, resulting in an increase in ABA and rapid fruit ripening in SlPLY9-OE compared with ABA-free and SlPYL9-RNAi plants and ABA-treated WT control fruit. The response in the SlPYL9-OE lines was further amplified by exogenous ABA application; however, in SlPYL9-RNAi fruit, reverse exogenous ABA did not rescue the delayed phenotype. In contrast, while treatment with NDGA significantly suppressed and delayed ripening in SlPYL9-RNAi fruit, reverse application did not block ripening of SlPYL9-OE fruit (Fig. 5A–U). Notably, prior to treatment, the ABA content of the OE fruit was high and that of the RNAi fruit low compared with the WT (Fig. 5V). The significant increase and decrease in ABA levels respectively caused by OE and RNAi of SlPYL9 might therefore be attributable to a feed-back effect on the number of available ABA signal molecules when the ABA signaling pathway is enhanced or repressed. At 4 d after ABA treatment, ABA content increased in all fruits (Fig. 5V). These results suggest that the ABA signaling process is enhanced and reduced by up- and down-regulation of SlPYL9, respectively, suggesting that SlPYL9 triggers early ABA signaling during onset of fruit ripening.

Fig. 5.

Effects of ABA and NDGA treatment on ripening onset in SlPYL9-OE/RNAi fruit at the mature green (MG) stage. A 0.5 ml volume of ABA (100 μM) or NDGA (200 μM) was slowly injected into 15 OE and 15 RNAi MG fruits (below the sepals) still attached to the plants. Control fruits were injected with distilled water. (A, E, I, M, Q) Untreated MG fruits. (B, F, J, N and R) Untreated fruits at 8 d. (C, G, L, P, T) Eight days after NDGA treatment; (D, H, K, O, S) 8 d after ABA treatment. (U) Breaker day. (V) ABA contents before and after ABA treatment in the control and treated fruit. *P<0.05; **P<0.01, t-test.

Effect of SlPYL9 on fruit ripening and quality

The obvious phenotypic change in the transgenic lines was related to fruit ripening. Fruits were therefore analysed for physiological and biochemical parameters (Fig. 6; Supplementary Fig. S6). As shown in Fig. 6A, fruit ripening was significantly accelerated by 6–7 d in both SlPYL9-OE lines compared with the WT. Compared with the breaking time (days from full bloom to fruit breaking) of the WT fruit (32 d), the breaking time was 25 d in OE line 3 and 26 d in OE line 11. In contrast, fruit ripening was delayed by 4 d in SlPYL9-RNAi7 and by 2 d in SlPYL9-RNAi 9/22 compared with the WT (Fig. 6A, B). These results suggest that SlPYL9 is involved in the regulation of fruit ripening, with a stronger effect of OE on ripening than of RNAi.

Fig. 6.

Fruit ripening in SlPYL9 overexpression and RNAi lines. (A) Fruit ripening in two SlPYL9-OE lines and three SlPYL9-RNAi lines compared with the WT. (B) Quantification of fruit ripening by days to reaching the breaker stage in WT and transgenic fruits. Thirty fruits per transgenic line were used to quantify ripening time. (C) The fruit shape index in WT and transgenic fruits during ripening. At least 30 fruits per genotype were evaluated. (D, E) The soluble solid content (D) and fruit firmness at harvest (E). (F) Change in ABA content (F). (G) Change in ethylene release. *P<0.05; **P<0.01, t-test.

In addition, SlPYL9-OE/RNAi also strongly affected fruit quality. As shown in Fig. 6E, fruit firmness of the WT at the red mature stage was lower than that of the RNAi fruit, but higher than that of the OE fruit. Moreover, the soluble solid content of the red mature fruit was 4.59%, 4.0%, and 5.2–5.4% in the WT, RNAi, and OE fruit, respectively (Fig. 6D). The fruit shape index (the ratio of fruit length to width) of the three SlPYL9-RNAi lines was higher than that of the WT and SlPYL9-OE fruit (Fig. 6C). Moreover, endogenous ABA accumulation and ethylene release were significantly advanced in the OE fruit and delayed in the RNAi fruit compared with the WT at each stage (Fig. 6F, G). The mesocarp of the three RNAi lines was thicker than that of the OE lines and WT, particularly RNAi-7, which had a mesocarp thickness of 3.69 mm, significantly larger than that of the WT (Fig. 7). In addition, the filled seed number was approximately 15–20 per fruit in the WT, but only 4–8 seeds were observed in the RNAi fruit and 0–6 seeds in the OE fruit (Fig. 7). Collectively, these results suggest that SlPYL9 is involved in ABA-mediated fruit development and overall quality in tomato.

Fig. 7.

Comparison of mesocarp thickness between the WT and transgenic fruits at different stages. MG, breaker and RR: transverse sections of the fruit in different stages. Mesocarp thickness: comparisons of mesocarp thickness in the OE/RNAi fruits. Thirty fruits per transgenic line were used to measure pulp thickness. **P<0.01, t-test. MG, mature green fruit; RR, ripe red fruit. (This figure is available in color at JXB online.)

OE and RNAi of SlPYL9 affect expression of ripening-related genes

The close relationship between the phenotype of the transgenic fruit and the ripening-related physiological parameters described above suggests a role of SlPYL9 in fruit ripening. To further understand SlPYL9-mediated regulation of fruit ripening, breaker fruit transcriptomes were analysed by RNA-Seq. The breaker stage represents the transition from development to ripening, and involves extensive reprogramming of fruit transcriptomes. In line with this, 1080 (OE3) and 1856 (OE11) genes were modulated by OE, and 3798 (RNAi7) genes were modulated by RNAi compared with the WT, suggesting that SlPYL9 extensively affects fruit cell transcriptomes (Fig. 8; Dataset 2; Kai et al. 2019). ABA levels are determined by biosynthesis and catabolism, which in turn are regulated by 9-cis-epoxycarotenoid dioxygenase (NCED) and ABA 8′-hydroxylase (CYP707A), respectively. There are four NCED and four CYP707A genes in the ABA metabolism pathway, of which NCED1 and CYP707A2, respectively, are the most important. As shown in Fig. 9A, the expression of SlNCED1 was up-regulated while the expression of SlCYP707A2 was down-regulated in the OE fruit during the ripening onset. It is worth noting that the increase of ABA accumulation caused by SlPYL9-OE during the mature green and pre-breaker stages (Fig. 6F) induces the ripening onset. However, when ABA accumulation reaches the maximum at the breaker stage, it in turn suppresses the fruit ripening. The evidence for this was provided by results from RNA-Seq, in which several key ripening regulators, including RIN (Vrebalov et al., 2002), CNR (Manning et al., 2006), TAGL1 (Vrebalov et al., 2009), FRUITFULL1 (FUL1) (Bemer et al., 2012), and SlHB1 (HD-Zip homeobox protein) (Lin et al., 2008), were all down-regulated in OE fruits during the breaker stage (Fig. 9E). In contrast, the transcript levels of these genes in RNAi fruits showed opposite results. Similarly, the transcript levels of most of the target genes encoding components of ethylene biosynthesis and signaling, carotenoid biosynthesis, and cell wall degradation, such as SlACS2, SlACS4 (encoding ACC synthase), SlACO1 (encoding ACC oxidase) (Fig. 9C), SlETR3 (encoding an ethylene receptor), SlCTR1 (encoding the constitutive triple response), APETALA2a (AP2a) (Karlova et al., 2011), and SlERF2 (encoding an ethylene response factor) (Fig. 9D), polygalacturonase (SlPG) and expansin (SlEXP6) (Fig. 9F), and phytoene synthetase (SlPSY1), phytoene dehydrogenase (SlPDS), carotene dehydrogenase (SlZDS), and lycopene β-cyclase (LYC) (Fig. 9G), were down-regulated in OE fruits but up-regulated in RNAi fruits. In addition, the expression of some genes encoding glucose metabolism (Fig. 9H) and fruit flavor (Fig. 9I) were significantly altered in transgenic fruits.

Fig. 8.

Gene expression responses to SlPYL9-OE/RNAi in the transgenic line. The Venn diagram represents genes with changed expression (log₂-fold change is >1 and <−1 with false detection rate <0.005) in OE/RNAi transgenic fruits compared with that of the WT fruits. (This figure is available in color at JXB online.)

Fig. 9.

The changes of gene expression involved in tomato fruit ripening for SlPYL9-OE/RNAi transgenic and WT fruits at the breaker stage. The genes related to ABA metabolism and signaling (A, B), ethylene synthesis and signaling (C, D), transcription factors (E), cell wall catabolism (F), lycopene biosynthesis (G), glucose metabolism (H), and fruit flavor (I). The data are based on the RNA-Seq of fruits at the breaker stage. Log₂-fold change is >1 and <−1 with false detection rate <0.005. The heat map was made with MEV4.9.0 software.

Moreover, the qRT-PCR results showed that, consistent with the change in ABA, expression of SlNCED1 and SlCYP707A2 was advanced in OE and delayed in RNAi fruits (see Supplementary Fig. S7A, B). Several ethylene-related genes were subsequently examined. Expression of SlACS2, SlACS4, SlACO1, SlCTR1, SlETR3, and SlERF2 was advanced in OE fruit, but delayed in RNAi fruit (Supplementary Fig. S7C–H). In addition, expression of genes related to lycopene biosynthesis, such as SlPSY1, SlPDS, and SlZDS (Supplementary Fig. S7L–N), and the cell wall catabolism pathway, such as SlPG and SlEXP1, and the xyloglucan endo-transglycosylase (SlXET16), showed similar changes (Supplementary Fig. S7I–K). These results suggest that the alteration in SlPYL9 expression is correlated with expression of ripening-related genes, thereby affecting the ripening process.

SlPYL9-RNAi affects flower abscission and fruit shape

The transgenic OE and RNAi lines showed multiple phenotypes in terms of vegetative growth and flower development. OE and RNAi of SlPYL9 led to decreased plant width, and the transgenic plants weighed approximately 25% less than the corresponding WT plants (see Supplementary Fig. S8). Flower abscission was also significantly different between the SlPYL9-RNAi lines and WT as well as between the SlPYL9-RNAi and SlPYL9-OE lines; however, no significant difference was observed between the WT and OE lines. Petal abscission in the three RNAi lines was delayed compared with the WT after full bloom and fruit setting (Fig. 10). The base of the petal was held tightly at the top of the young fruit, causing pointed/oblong fruit in RNAi-7/9 (Fig. 10A, B), while in RNAi-22, the fast-growing fruits were gourd shaped (Fig. 10C). With the expansion of young fruits, the base of the petal was at last broken followed by abscission. Accordingly, all RNAi-7/9 fruits were pointed/oblong, while approximately 40–60% of the RNAi-22 fruits were gourd shaped, and abscission of the flower stigma was relatively late compared with the WT. The ABA accumulation in RNAi transgenic flowers was lower than that of the WT flower (Fig. 10D). Taken together, these results suggest that SlPYL9 is also involved in petal abscission via the regulation of ABA level, thereby affecting fruit shape.

Fig. 10.

Phenotypes of petal abscission in the three SlPYL9-RNAi lines. (A–C) Petal abscission in SlPYL9-RNAi 7 (A), SlPYL9-RNAi 9 (B) and SlPYL9-RNAi 22 (C). (D) ABA content in the whole flower except for sepals. Thirty flowers per transgenic line were used to measure the ABA content. *P<0.05, t-test.

Discussion

The ABA receptor PYR/PYL/RCAR or PYLs in Micro-Tom tomato

Using Micro-Tom tomato as test material revealed two points worth noting. The first is that this variety possesses only a small PYL family. The transcriptome data (Dataset 1; Kai et al. 2019) revealed expression of only 10 SlPYL genes in breaker fruit, the peak point during fruit development. This result suggests that the number of PYL family receptors is less than in Arabidopsis and other tomato varieties. Similarly, in a recent paper, Brachypodium was also found to have only a small number of ABA receptors (Pri-Tal et al., 2017). It is generally believed that Arabidopsis shows functional redundancy in ABA receptors, with the complex pyr1–pyl1–pyl2–pyl4 quadruple mutant leading to broad spectrum ABA hyposensitivity (Park et al., 2009; Okamoto et al., 2013). Thus, this suggests that single receptor loss-of-function would result in a very weak ABA-insensitive phenotype. However, genetic and biochemical analyses of ABA signaling components in tomato and Arabidopsis revealed that despite high conservation, the extent of receptor redundancy differs between the two species. Herein, this raises the question as to whether each PYL shares more work in the case of a small ABA receptor subset. In this study, both qRT-PCR and RNA-Seq analyses revealed very low transcript levels of SlPYL9; however, manipulation of this gene alone revealed significant differences in ABA signaling and obvious changes in the phenotype of transgenic fruits/plants. These findings suggest, therefore, that OE/RNAi of a single SlPYL9 results in a higher phenotypic marginal value, with each PYL sharing more work in the small SlPYL subset in Micro-Tom tomato.

Second, use of the Micro-Tom variety as experimental material highlighted certain issues that need to be addressed, especially in the context of hormone studies. Micro-Tom harbors a brassinosteroid deficiency mutation (the Dwarf gene involved in BA synthesis), which increases gibberellic acid (GA) sensitivity. However, in this study, our findings did not appear affected. There are two possible reasons for this; first, since SlPYL9 expression was specifically manipulated by OE/RNAi, expression of the genes related directly to SlPYL9, such as SlPP2C1/2/9 and SlRnRK2.8 (Fig. 4) and their downstream counterparts, were also affected. Second, since the concentrations of brassinosteroid and gibberellin were very low during fruit ripening, the contents and expression of their genes were largely unaffected by SlPYL9-OE/RNAi (Datasets 1–3; Kai et al. 2019). However, despite the involvement of brassinosteroid and gibberellin in plant development, no significant changes were observed between the WT and transgenic lines at the seedling stage. That is, the transgenic phenotypes were not exacerbated by the additional hormonal issues.

SlPYL9 is involved in the regulation of ABA signaling during fruit ripening

In ABA signaling, the interaction between the ABA receptor PYL and PP2C is the key step that activates the downstream signaling genes to evoke ABA responses (Fujii et al., 2009). Tomato possesses 14 ABA receptors, among them, SlPYL9, if it is an ABA receptor, would be expected to interact with and inhibit SlPP2Cs, thereby activating ABA signaling. This hypothesis was verified by the SlPYL9 inhibition of SlPP2C2/6 activity (Fig. 2B, C) and by the physical interaction between the SlPYL9 protein and SlPP2C2/3/4/5 in an ABA-dependent matter (Fig. 3A, B). These experiments show that SlPYL9 acts as a potent inhibitor of phosphatase activity in the presence of ABA. We also evaluated sensitivity of the transgenic lines to ABA. SlPYL9-OE resulted in ABA hypersensitization, while silencing of SlPYL9 had the opposite results (see Supplementary Figs S4, S5, S9). These findings are also in line with previous reports on ABA receptors (Saez et al., 2006; Finkelstein, 2013; Zhao et al., 2014, 2016).

We also investigated the sensitivity of the transgenic fruit to ABA during ripening onset. When MG fruits were treated with ABA, rapid ripening of OE fruit was amplified due to both the increased ABA level and the enhanced ABA signaling caused by SlPYL9-OE. In contrast, NDGA, an inhibitor of the NCED enzyme (Creelman et al., 1992), was used to fully block ABA accumulation. As shown in Fig. 5, NDGA treatment delayed the ripening of PYL9-RNAi fruits compared with the untreated PYL9-RNAi fruits, and caused decreased ABA level. ABA application triggered a similar response in both transgenic and WT fruits (Fig. 5V), suggesting that the entire ABA response is mediated by additional receptors. Overall, these results suggest that receptor redundancy does exist in Micro-Tom tomato.

ABA is a basic hormone for plant development, and its synthesis and signaling are controlled stringently by the related developmental program. The ABA levels and signaling have to be moderate during development. In the breaker stage, the expression of SlPYL9 was in a rapid rising phase, and the artificial enhancement or suppression of this gene will cause the stress response of co-receptor PP2C to prevent excessive change of SlPYL9 expression. After the stress reaction, PP2C will return to its normal response level. We noticed that the expression levels of several SlPP2Cs showed an upward trend in both RNAi and OE fruits (Fig. 4B). This phenomenon can be considered as the stress expression of PP2C, which is partially detected in the sample. However, from the view of biological effects, the OE and RNAi fruits showed opposite ripening phenotypes, and the transcriptome data (Fig. 8) were consistent with the phenotype of the fruits, i.e. the expression of downstream ripening-related genes went in the opposite directions in response to ABA signaling in OE and RNAi fruits. In addition, this is related to the specificity of RNAi, which may be on the low side. Further experiments are needed to explore this question accurately. Next, among SnRK2 genes, only SnRK2.8 was significantly down-regulated while others showed no change in the OE fruits; however, SnRK2.8 was up-regulated together with other SnRK2s in the RNAi fruits. This result suggested that the low/high SnRK2.8 activity observed in the PYL9-OE/PYL9-RNAi lines might be a direct consequence of PP2C–SnRK2 interactions. As SnRK2 is at the end of the PYL–PP2C–SnRK2 signal chain, the results above indicated that at the breaker stage, ABA signaling was reduced in OE fruits but enhanced in RNAi fruits (Fig. 4C).

Expression profile of SlPYL9 during fruit ripening

Tomato is a typical climacteric fruit, ripening and quality of which can be modified by SlPYL9-mediated signaling. SlPYL9 plays a role in fruit ripening according to several pieces of evidence. (i) SlPYL9 is expressed in fruits during development and ripening. Although the expression level of SlPYL9 is not very high, it sufficiently elicits the response of downstream ripening-related genes, because as PYL9 is an ABA receptor, a small change is dramatically amplified, which can lead to a large response. (ii) During the key transition from MG fruit to breaker fruit, the expression levels of SlPYL9 increase several times, and its expression pattern is in parallel with the ABA accumulation and SlNCED1 expression, suggesting that it is involved in the regulation of ABA-mediated fruit ripening. Among 14 tomato PYLs, only the change of PYL1/4/9 is consistent with the change of ABA accumulation in fruit during development and ripening. (iii) Furthermore, the overexpression and interference of PYL9 alter the fruit ripening time and fruit quality in tomato. These results demonstrate that the SlPYL9 plays a role in the regulation of fruit ripening.

SlPYL9 is involved in the regulation of ABA-mediated fruit ripening in a complicated manner. It is found that compared with the WT fruits, SlPYL9-OE results in increased ABA accumulation during the mature green and pre-breaker stages (Fig. 6F), which leads to induced ethylene release at the same time (Fig. 6G), thereby triggering ripening onset. Thus, the ripening onset of the OE fruit is earlier than that of the WT fruit. Following this, when ABA accumulation reaches the maximum at the breaker stage, it in turn suppresses ethylene production (Fig. 6G), which supports negative cross-talk between the two hormones (Tieman et al., 2001; Pech et al., 2008; Rodriguez et al., 2010). The decreased ethylene impacts fruit coloring and cell wall degradation. During B+3 to B+5 stages, in OE fruits, the coloring is slower and the firmness is higher compared with the WT fruits. However, during the ripe stage from B+7 to B+10, the OE fruits rapidly soften, and the color is orange–red instead of red as in the WT fruits (Fig. 6A), which is caused by a greater β-carotene and lower lycopene content compared with the WT fruits (see Supplementary Fig. S5). Expression of carotenoid biosynthesis genes, including the gene encoding ethylene-regulated phytoene synthase 1 (PSY1), a rate-limiting enzyme in carotenoid synthesis (Fraser et al., 2002; Martel et al., 2011), shows a significant decrease in OE fruits but an increase in RNAi fruits at the onset of fruit ripening (Fig. 9G). Also, SlPYL9 negatively regulates genes of carotenoid isomerase (CRTISO) (Isaacson et al. 2002) and LYC (Solyc06g074240), while it positively regulates genes of lycopene β-cyclase (LYC-B) (Dataset 1; Kai et al. 2019). The decreased lycopene is due to an increased activity of lycopene β-cyclase, which channels fewer metabolites into the lycopene pathway, thereby increasing carbon flux to β-carotene and ABA. This is consistent with our results in the previous reports (Sun et al. 2012a,b; Sun et al. 2017). These findings suggest that the genes related to ethylene release, pigment metabolism, and cell wall degradation are all regulated by SlPYL9-mediated ABA signaling during ripening (Fig. 9).

Role of SlPYL9 in the regulation of ABA-mediated petal abscission

OE and RNAi of SlPYL9 resulted in polytropic phenotypes during flower abscission. Jasmonate, indole-3-acetic acid (IAA) and auxin are all known to play important roles in flower development from anthesis and sex determination to flower abscission (Ludwig-Müller et al., 2009; Yuan and Zhang, 2015). However, whether the ABA receptor PYL is also involved in flower abscission remains unclear. To confirm this, we investigated transgenic flowers during anthesis. In the three SlPYL9-RNAi lines, 30–50% of the petals did not abscise normally compared with the WT. This was particularly the case in RNAi line 22 (Fig. 10C), in which the base of the petal held firmly to the top of the young fruit, resulting in a gourd-shaped fruit and thereby affecting the fruit shape index. These results suggest that petal abscission is determined by a dynamic balance of hormones (GA, IAA, and ABA), modulated by SlPYL9-mediated ABA signaling during anthesis. Furthermore, they also suggest that the interaction between genes is characterized by a more variable profile related to flower/fruit tissue-specific modulation, warranting further characterization.

In conclusion, the results of this study suggest that SlPYL9 is an ABA receptor that positively regulates ABA signaling and fruit ripening. In line with this, SlPYL9-OE/RNAi plants showed typical ABA hypersensitive/hyposensitive phenotypes. Moreover, SlPYL9-OE significantly accelerated fruit ripening by 5–7 d, while SlPYL9-RNAi delayed fruit ripening, enhanced mesocarp thickness, and hindered petal abscission. These results suggest that SlPYL9 is involved in ABA signaling, participating in the regulation of fruit ripening in Micro-Tom tomato. These findings give us further insight into the relationship between ABA receptors and flower/fruit development.

Supplementary data

Supplementary data are available at JXB online.

Fig. S1. Phylogenetic analysis of tomato ABA receptor PYLs and Arabidopsis AtPYLs.

Fig. S2. Expression levels of SlPYL genes in fruit during development.

Fig. S3. Construction of SlPYL9-OE/RNAi expression vectors.

Fig. S4. Seed germination and primary root growth in WT and SlPYL9-OE/RNAi plants.

Fig. S5. Comparison of drought resistance in SlPYL9-OE/RNAi lines and WT plants.

Fig. S6. Lycopene, beta-carotene and total carotenoid content in ripe fruits.

Fig. S7. Expression of ripening-related genes in the WT and SlPYL9-OE/RNAi fruits during development.

Fig. S8. Comparison of internode, plant fresh weight and plant sizes between WT and transgenic plants.

Fig. S9. Effects of dehydration stress on WT and transgenic fruit.

Table S1. Specific primers used for gene amplification for SlPYL9-OE/RNAi plasmid construction.

Table S2. Specific primer sequences used for real-time quantitative PCR analysis.

Table S3. Quantitative PCR primers used to analyse the expression of ripening-related marker genes in SlPYL9-OE and RNAi lines.

Table S4. Specific primer sequences used for subcellular localization assay of SlPYL9.

Table S5. Specific primer sequences used for SlPYL9 and SlPP2Cs in a bimolecular fluorescence complementation (BiFC) assay.

Table S6. Specific primers used for SlPYL9 expression in situ hybridization.

Data deposition

The following data are available at Dryad Data Repository (https://doi.org/10.5061/dryad.5q449h5Kai et al. 2019).

Dataset 1. The expression of fruit ripening-related genes in fruits at the breaker stage detected by RNA-Seq.

Dataset 2. Genes expressed differentially in the WT and transgenic fruits at the breaker stage detected by RNA-Seq.

Dataset 3. Total gene numbers in the WT and transgenic fruits at the breaker stage detected by RNA-Seq.